03:48
<annevk>
roc: heh, you are!
06:49
<foolip>
Hixie: pong, I've moved it up on my todo list
06:56
<annevk>
I wonder if DreamHost has MySQL 5.5 now so I can do this: http://mathiasbynens.be/notes/mysql-utf8mb4
06:56
<annevk>
Using an emoji required a character reference :/
07:05
<foolip>
annevk: the new exitFullscreen() doesn't actually sync fullscreenchange events with the resize
07:06
<foolip>
like requestFullscreen(), the "as part of the next animation frame task" needs to be the first step after "resize"
07:07
<annevk>
foolip: I thought resize -> change state for request meant change state -> resize for exit, but yeah, I guess
07:07
<annevk>
foolip: there's the thing though that you don't always resize
07:07
<annevk>
foolip: and you only resize one you you realize the stack is empty
07:08
<annevk>
foolip: and you don't know about that until you start popping it
07:10
<foolip>
yes, it'll be a bit complicated
07:10
<foolip>
do you have an idea, or should I think hard?
07:10
<annevk>
my idea was to do the resize afterwards :)
07:11
<foolip>
that'll lead to flicker and scripts seeing an inconsistent state though
07:11
<annevk>
open a bug and I can give it another go I suppose
07:11
<foolip>
https://www.w3.org/Bugs/Public/show_bug.cgi?id=26440 is still open, I'll comment
07:12
<annevk>
I guess we can figure out whether the stack will be empty upfront
07:12
<annevk>
although we'll also have to determine it for all ancestor browsing contexts, lots of cross-process shit
07:12
<annevk>
but I decided not to worry about those things anymore, at least not too hard
07:13
<foolip>
right, I saw the note :)
07:14
<foolip>
right now there is no browser with out-of-process iframes I think, so no one will notice if the model is broken until later
07:14
<foolip>
yay?
07:15
<foolip>
synchronous IPC does exist, so maybe it'll be fine
07:15
<annevk>
the Servo crowd will not like me
07:15
<foolip>
so ask them to review this stuff!
07:16
<foolip>
Presto was single-process single-threaded, so what do I know? :)
07:16
<annevk>
Yeah I should I guess, I keep forgetting who I should ping, maybe SimonSapin can help out
07:16
<annevk>
Presto had its own magic
07:17
<foolip>
yeah...
07:21
<mathiasbynens>
annevk: $ mysql --version
07:21
<mathiasbynens>
mysql Ver 14.14 Distrib 5.1.73, for debian-linux-gnu (x86_64) using readline 6.1
07:21
<mathiasbynens>
:(
07:23
<annevk>
sad times
07:24
<zcorpan>
Hixie: if the preloader can't evaluate a mq it doesn't preload anything. that was a requirement in the spec before the HTML merge
07:30
<annevk>
mathiasbynens: did you ever research why utf8 was so bad in the first place?
07:33
<foolip>
annevk: https://www.w3.org/Bugs/Public/show_bug.cgi?id=26440#c11
07:42
<SimonSapin>
annevk: ?
07:46
<SimonSapin>
annevk: what about Servo / how can I help?
07:47
<mathiasbynens>
annevk: nope, but my guess would be incompetence
07:48
<SimonSapin>
annevk: Do you mean not-really-utf8-in-mysql? Apparently at the time when Unicode was only the BMP, UTF-8 only needed up to 3 bytes per code point
07:58
<annevk>
SimonSapin: I'd like whoever works on out-of-process iframes to look at Fullscreen
08:00
<annevk>
foolip: currently step 3.4.3 has a loop when popping that you miss
08:00
<SimonSapin>
ok. Any point specifically?
08:01
<annevk>
SimonSapin: I guess whereever the source code says "cross-process"
08:01
<annevk>
SimonSapin: it's whenever frame boundaries are crossed so any talk about browsing context is suspect
08:02
<SimonSapin>
ok
08:08
<foolip>
annevk: I thought you would say that. Why not just remove elements from the middle of the stack when they're removed from the document?
08:09
<annevk>
foolip: makes sense, I wonder why I forgot about that, I'll clean that up first
08:22
<annevk>
foolip: I thought I wrote "ordered set" in the spec, fixing that now
08:24
<foolip>
annevk: thanks for fixing things so promptly!
08:24
<foolip>
let's hope there's enough time to implement and ship it with the next release
09:34
<mathiasbynens>
SimonSapin: are you sure MySQL’s `utf8` was implemented way back then?
09:35
<SimonSapin>
no idea, just a guess
09:36
<mathiasbynens>
“Up to and including TUS 3.0.1, characters were assigned only in the BMP. In TUS 3.1, characters were assigned in the Supplementary Planes for the first time, in Planes 1, 2 and 14.”
09:37
<mathiasbynens>
Unicode 3.1.0 == March 2001
09:39
<mathiasbynens>
can’t seem to find when MySQL added its `utf8` implementation though
09:44
<annevk>
but it was known from Unicode 2 that could happen
09:44
<annevk>
but yeah, makes sense I suppose
09:44
<annevk>
but that it takes a decade to fix...
09:46
<mathiasbynens>
https://web.archive.org/web/20040612104116/http://dev.mysql.com/doc/mysql/en/Charset-Unicode.html
09:46
<mathiasbynens>
“As of MySQL version 4.1, there are two new character sets for storing Unicode data: `ucs2` [… and] `utf8`”
09:46
<mathiasbynens>
there we go
09:47
<mathiasbynens>
MySQL 4.1 was released in October 2004
09:47
<mathiasbynens>
so yeah, incompetence
09:48
<annevk>
mounir_: should prolly run DOMQuatern past public-script-coord too?
09:48
<annevk>
wow
09:48
<annevk>
no utf-16?
09:49
<annevk>
because that would sort of explain utf8, they just mapped to ucs2
10:07
<mounir_>
annevk: oh, they do geometry stuff?
10:29
<Ms2ger>
annevk, happy birthday
10:29
<mounir_>
annevk: oh indeed
10:29
<mounir_>
annevk: you can no longer call me "old man" \o/
10:30
<jgraham>
mounir_: I wouldn't bet on that
10:31
<mounir_>
jgraham: I wouldn't bet that he will stop doing it
10:33
<jgraham>
mounir_: We seem to be in agreement. Suddenly everything seems strange in the world.
10:35
<mathiasbynens>
annevk: oh, look at the date! happy birthday
10:47
<foolip>
annevk: why does parentDocs need to be kept across the async boundary, can't resize be set to true before the async section?
10:47
<foolip>
probably some raciness I'm not seeing
10:51
<foolip>
annevk: missing "a" in "If exitDocs's first document has browsing context container"
11:32
<annevk>
foolip: oh, we could do that the other way around
11:33
<annevk>
foolip: didn't really think about that
14:17
<smaug____>
TabAtkins: could you perhaps file a bug to get chrome too to warn about use of sync xhr ?
14:20
<TabAtkins>
Sure.
14:20
<TabAtkins>
Firefox warns; doors anyone else?
14:21
<Ms2ger>
Doors?
15:02
<TabAtkins>
Ms2ger: Bah, phone keyboards.
16:21
<Hixie>
zcorpan: isn't that just going to result in a bunch of slow-loading pages? i thought the whole point here was to carefully design the feature so that it could be pre-loaded.
20:16
<Hixie>
can someone load http://www.hixie.ch/tests/adhoc/html/video/001.html in a recent IE and tell me how noisy it gets? :-)
20:34
<Hixie>
GPHemsley: dude someone cut off the corner of the file in the icon for mimesniff :-P
20:34
<Hixie>
GPHemsley: either that, or they folded it backwards from the usual way it's folded :-P
20:40
<miketaylr>
Hixie: this is what i see in my vmware/IE11 https://cloudup.com/cTF_o3DUg-8
20:42
<Hixie>
wow
20:42
<Hixie>
thanks
20:42
<Hixie>
miketaylr: can you check what version of IE11 that is?
20:42
<Hixie>
newer or older than v11.0.7 ?
20:42
<miketaylr>
one sec
20:43
<miketaylr>
Hixie: https://cloudup.com/cdicl-Y0S0h
20:43
<miketaylr>
(sorry, can't select the version to just paste in here)
20:43
<miketaylr>
so i guess newer
20:44
<Hixie>
weird
20:44
<Hixie>
thanks!
20:45
<Hixie>
(https://www.w3.org/Bugs/Public/show_bug.cgi?id=11984)
20:47
<Hixie>
GPHemsley: i'm going to be using the "rules for sniffing audio and video specifically", fyi
21:32
<Hixie>
annevk: http://dom.spec.whatwg.org/#dom-document-createevent points to https://dvcs.w3.org/hg/d4e/raw-file/tip/source_respec.htm for UIEvent but that spec doesn't seem to define it
21:32
<Hixie>
annevk: what's the story with that?
21:58
<Hixie>
zcorpan: btw you have a dangling <code>srcset</code>
22:11
<Hixie>
is there a quick trick for replacing an element with its children, in JS/DOM? or do i just have to do it manually
22:35
<gsnedders>
Hixie: how evil? el.outerHTML = el.innerHTML should work, no?
22:35
<Hixie>
that makes scripts re-execute
22:35
<Hixie>
i went with just making a docfrag of the contents, then doing a replacechild()
22:35
<gsnedders>
yeah, if you have scripts don't think you can do better
22:41
<Hixie>
wtf
22:42
<Hixie>
http://www.buzzsugar.com/ uses <article>, <time>, <header>, <nav>, and all kinds of other newfangled features
22:42
<Hixie>
but they can't use <h1>-<h6>?!
22:53
<gsnedders>
wat
22:53
<gsnedders>
"can't"?
22:53
<Hixie>
apparently
22:53
<Hixie>
there isn't a single heading on that page
22:53
<Hixie>
it has an elaborate outline, but no headings
22:54
<gsnedders>
Maybe you should try using the Elaborate Description of the Outlining Algorithm?
22:54
<gsnedders>
I think it's in Appendix E?